Holger Ebel wrote:
>
> Hi,
>
> i expect that the following should work (under 3.22.22):
>
> select distinct DATE_FORMAT(datum, '%d.%m.%Y') from Info_lnk order by
> datum; (where datum is a timestamp(14) field)
>
> and give me a result with the unique data like:
>
> +----------+
> |10.10.1999|
> |11.10.1999|
> |12.10.1999|
> .....
>
> but instead of this, i get:
>
> +-------------------------------+
> |10.10.1999 |
> |10.10.1999 |
> |11.10.1999 |
> |11.10.1999 |
> |11.10.1999 |
> |12.10.1999 |
> |12.10.1999 |
> .....
> (takes 1.02 s with -->>3555<<-- rows)
>
> using select DATE_FORMAT(datum, '%d.%m.%Y') from Info_lnk order by
> datum
>
> +-------------------------------+
> |10.10.1999 |
> |10.10.1999 |
> |11.10.1999 |
> |11.10.1999 |
> |11.10.1999 |
> |12.10.1999 |
> |12.10.1999 |
> .....
>
> (takes 0.61s with -->>7370 rows<<--)!!!
>
> The same thing happens, when i use
>
> select distinct left(datum, 8) from Info_lnk order by
> datum;
>
> so, is this a bug or a feature?
>
> cu
>
> Holger
Hi Holger
This is not a bug!
Because you use an ORDER BY datum, datum is implicitely added to the SELECT result.
With DISTINCT you only eliminate those entries with identical datum.
datum is a timestamp, so it must be equal down to seconds!
What you search for is:
SELECT
DATE_FORMAT(datum, '%d.%m.%Y') AS fdatum
FROM
Info_lnk
GROUP BY
fdatum
Tschau
Christian
